Anthropic, Nvidia, Stock Crash, Workday Rising, Greed, Politics, Business: Making Sense of It All.

November 21, 2026 • 00:27:33
Anthropic, Nvidia, Stock Crash, Workday Rising, Greed, Politics, Business: Making Sense of It All.
The Josh Bersin Company
Anthropic, Nvidia, Stock Crash, Workday Rising, Greed, Politics, Business: Making Sense of It All.

Nov 21 2026 | 00:27:33

/

Show Notes

I just spent two long weeks in the Middle East, Geneva, and Barcelona meeting with hundreds of companies (manufacturing, tech, airlines, defense, financial services) including a week at Workday Rising. And while I was gone Nvidia crashed, Anthropic went public, Trump went to China, Outlook and the US air space went down, and the AI-stocks took a nosedive.

In today’s recap I discuss the role of greed, the immaturity of this massive market, and what pragmatic, solution-oriented companies can do to prepare for the AI transformations ahead. As I describe, we’re dealing with a very immature technology and immature market. So rather than rush into major decisions, it’s ok to take a step back and look at the AI tsunami as an opportunity for strategic problem-solving.

I also saw Ethan Mollick present to a series of executives and saw first-hand how “hyping AI” can create a frenzy of FOMO (fear of missing out).

My message, as you’ll hear, is to take a deep breath and recognize that we’re still in the early days of this revolution, so to speak, and a culture of problem-solving and organizational growth is the best way to proceed. So I step through HR2030® and the four big use-cases for AI in HR and business.

And no, despite Dario’s ongoing claims, we are not going to wipe out the job market.

Read our new book Superpowered: How AI Unlocks the Superpower of People, Leadership, and Teams, to really understand what’s happening. This book is written for you as a business person, HR leader, or professional.

And if you want to stay current on all our research, get Galileo®, the AI Agent for everything HR.

Chapters

View Full Transcript

Episode Transcript

[00:00:00] Good morning, everyone. I'm back from two weeks in the road and three or four days at workday rising in Barcelona. And while I was gone, the entire AI industry imploded and exploded at the same time. And we see greed, power, strategy, value, marketplace, economics, politics, all coming together. So for those of you that don't follow all of the details of this, going to give you some perspectives on the big picture here. So, number one, of course, Nvidia has crashed. Not only, you know, for the Jensen Huang issue and a big New York Times article on Jensen Huang's poor behavior, which I believe has a huge impact on the success of a company. [00:00:47] I don't believe a CEO who lies or cheats or misinterprets what we hear in the stock market is going to build a sustainable company over the long run. Even though he did build the company from scratch, it probably is a good thing that he's gone good for Nvidia. But the company's imploding because there's lots of evidence that most or a large percentage of their revenue comes from customers who were financed by Nvidia and didn't really have the money to pay for the chips, that they sold them at a very high price. And some of the statistics seem to say that 50, 60, 70% of the IT costs in a data center go to them, to one company. So that's going down. Anthropic went public. Stocks started at 83, went up to about 130 for a few minutes, is back down in the 80s again. I don't know what it'll do next week because we saw essentially a flipping process where the early investors and maybe some of the employees flipped their stock immediately on all the institutional investors and the retail investors took it away from them, gave them all a profit, and are now stuck with a company that doesn't have any prospect of going profitable for a long, long time. And I would personally guess that Anthropic stock is going to go down in the near term before it goes up. But that created a lot of animosity and excitement, excitement on Wall street, of course, animosity amongst regular Americans that how did these guys make so much money in four and a half years when I can't pay for my groceries? [00:02:26] And that's a real issue. It affected our US Political system. It affects what people buy. It affects the products and services that you sell. [00:02:35] So we have that going on. Then we have this weird situation in China where on the one hand, we're not supposed to be sending them proprietary technology because we believe we're in a long term economic competition with them, but actually we are. It's okay. Harold Lutnick thinks it's fine. So the Commerce Department both says one thing and maybe it does something different. We don't really know why this is happening. We don't really know what's happening. So the new Congress isn't going to get underneath that. Then another topic that came up in Europe with Workday, which I'll talk about in a minute, is the sovereignty of AI. Sovereignty means do I have control over this? [00:03:15] And. And one of the most fascinatingly distressing thing that I've seen in my career was Elon Musk taking personal advantage of Starlink and deciding on his own what level of service we are going to give to Ukraine. And most people in Europe at least, and certainly everywhere I've met, don't like that. I don't think most Americans like it either that this one man seems to have control over such an important infrastructure. And, and there are a band of countries now teaming together to compete with that. And I think it's going to hurt Starlink along over many years, maybe not in their turn. But the topic is, well, if he can do that and he can run Doge and he can look at all our personal information and he can affect the government to give him billions and billions of dollars of revenue for Starlink and SpaceX or space as whatever you want to call it, Space sa, I guess he's renaming it. Why would I do business with a provider whose data centers are in the United States, where there's so much corruption that we don't know who's looking at our data? Because ultimately, in the big corporate market for AI, this is a proprietary tool used by you for your company. And you don't want your provider sniffing around, inspecting, leaking, sending your information somewhere else. So Workday created a whole company with only European nationals to run Workday in Europe, so that European countries would feel comfortable with the sovereignty of this system. I can't see any reason why this won't get worse and worse and worse, at least for the next few years. [00:04:55] So you're going to have to deal with that as a buyer. And then those of you that are providers are going to have to figure out which platforms you use and how you do business around the world to deal with this issue of sovereignty. Now, in the long list of topics, let's talk about AI safety. So before Anthropic listed publicly, I was sort of watching this actually, because I was in my hotel room. Dario and his sister were on the NASDAQ feed giving us all a speech about AI safety and how important it is to be aware of it. I don't know what to make of this. I think the big vendors like OpenAI and Anthropic have way overplayed their hand here. I think what they seem to be doing is pushing a narrative that this stuff isn't safe and therefore they, as the biggest providers should be given the authority to negotiate and work with the government on, on what the safety policy should be. I was reading a very interesting book about the history of the automobile. [00:05:59] In the automobile. The executive that ran Studebaker in the, I think it was in the twenties had exactly the same situation where before we had seat belts and crumpled glass and you know, lots of safety things in our cars. A lot of people were killed in cars. I remember when I was a young man in the 50s and 60s, cars weren't that safe and the car manufacturers really didn't do anything about it. Nor of course did the cigarette manufacturers and people who made a lot of things. So the government set up, you know, a lot of infrastructure to do testing, to come up with new technologies to study seat belts. I remember when the first seat belts were created. My dad had a Chevrolet Impala which was this beautiful red, very long convertible and he had to go have the car retrofitted with seat belts. They had to go in and build and attach the seat belts to the chassis in order for us to drive the car. Because Sometime in the 60s, I guess the seat belt laws were passed that wasn't done by some private company. [00:07:06] So on the one hand all of us are nervous about it because I heard it from a lot of people when I was in Europe. But on the other hand, we, we don't know what to believe because I don't want anthropic deciding what's safe. I don't want OpenAI deciding what's safe. I don't want Google deciding what's safe either. Those are profit making enterprises. So. And the government wants nothing to do with this. Donald Trump thinks this is a race and there's no way he's going to slow it down with stupid stuff like regulation. Okay? So we have that. And at the same time that's going on, we found out that the United States is now a country with measles. So much for government regulation or lack thereof there. So we have that. And you know, I don't know where this is going to go, but I do believe we will end up with an AI Safety or an AI institution in the government. We really need it. We need a lot of scientists and specialists in the federal government acting on our behalf to hold this industry accountable. And there's plenty of people that know what AI is and there's plenty of people understand how it works. These two big companies are definitely not in any way the only ones who understand this stuff. And then there's this interesting issue that all these data centers went down while I was gone. Outlook went down, the FAA went down, JPMorgan Chase went down, Workday went down, a whole bunch of banks went down. [00:08:20] And no one knows why, but it seems pretty clear that something around AI has to do with this because the agents are so good at finding exploits. In fact, Satya Nadella made a comment on DealBook. You might want to listen to his speech there. I'll say, I'll put that in the notes that the reason there are more outages is because it is easier for agents to make mistakes and you have to fix them much more quickly. And we are not used to that. We're used to taking time to fix cyber mistakes. So we don't really know what that's all about either. [00:08:54] And then there's Microsoft. So let me talk about Microsoft. By the way, it's interesting to me that I really think, and I talked a lot about this with the clients in workday, that the big IT vendors are going to be the safer providers of AI. Leaving OpenAI and Anthropic aside for a minute, Microsoft Workday, if you trust the vendor you have now, you're probably going to want to use them for your AI solutions because they have, number one, they have the global reach, they have the staying power to outlast the flurry of startups. And they have the data and the guardrails in their core systems to help you train your AI agents. If you use Microsoft, all your agents are trained by the Microsoft graph. If you use Workday, all your agents are trained by workday, coupled with Galileo, which will be embedded into Workday. So whereas if you start from scratch, you're doing the training, you're doing the data import, you're doing the data management, and you have to figure out how these things work together. And while I was at workday, we had a lot of conversations, a lot of companies about the end of jobs risk, where to start, how to cost justify this stuff. [00:10:06] And so let me give you some high level summaries of what I talked about there. Okay? So number one, every company needs an AI harness and an AI platform at its core. Just like you need Microsoft Outlook or an email system or a scheduling system for meetings, you need, you need something. Microsoft's tools may be a little funky to use, but they're really building the whole system. They're building everything you need in one platform. And I think a good 70% of you are probably going to end up standardizing on the Copilot. That's the reason we now have the Galileo platform integrated and optimized in the Copilot. So if you're a copilot shop, please talk to us. And the HCM side. I'm a huge fan of Workday, and not only because I'm working with them, but I really think they're taking a measured and strategic view of this. And the SANA platform is in a class by itself, in my opinion, for many things. You're going to read about a lot of these things in the 2027 report that I'm working on right now. So I think Workday is a very credible, strong AI platform. SAP is also doing an amazing job of building an AI platform. On top of the multimodule, highly interconnected nature of SAP, it's not as clean or elegant or integrated as Workday. They don't have the AI lab capabilities that Workday does in Sana. Sana is an AI lab, not just an AI product company. [00:11:32] I think they're going to be a major player here. I can't speak to Oracle, but I will say that I talked to a bunch of Oracle customers in the Middle East. None of them are happy. I had a number of customers say to me, we wish we could get off Oracle because it doesn't do what they said it does. I'm going to get in trouble for saying this on the podcast, but I would like to talk more with Oracle about what they're up to and why they keep overselling stuff, but that seems to be part of their DNA. Other vendors like ukg, Dayforce, adp, hibob are going in the same direction. So this is going towards your incumbent vendors. I know that is not the SaaS apocalypse hypothesis, it's the opposite. [00:12:14] But I'm more than sure that that's where the market's going to go. Second question Are we going to eliminate all our jobs? What's going to happen to the job market? So let me spend a few minutes on that. I had a whole Q and A on this on stage with an executive from Accenture. We are not going to eliminate all our jobs. The job market is going to get tighter. In fact, we're going to have a labor shortage. You can read about that on the articles on our website. You need to take good care of your people because they are going to be managing the agents. The agents aren't going to be managing them. The agents don't build themselves, the agents don't tune themselves, the agents don't align themselves. Humans have to decide what the agent architecture is like. Humans have to take care of the agents. I have not talked to a customer service agent at all ever in my, you know, so far, last two to three years that I like. So this is not as easy as it sounds. [00:13:07] The vendors do so much work, but the bulk of the work is done by you. So whatever agents you build and you can use the HR 2030 framework to help you with that, you're going to be doing that work. So that's going to take people. Second thing relative to jobs is this task intelligence market. [00:13:24] So a couple of vendors, SAP Workday and some others are now promoting this idea of job task analysis using AI. And you sit down and you look at all your job titles in the company, both Workday Workday has a tool for this SAP bot, a vendor trying to get into this market. And you look at your job titles and the system looks up the job titles in a bunch of databases and says, oh, that job's 30% automatable. Okay, that's a nice piece of information. What am I going to do with that information? It's helpful because you can see where the automation opportunities are. We've been through this for a couple of years. We worked with Rejig, we've worked with Lightcast on this. [00:14:05] I think having an analysis of the task to job relationship in your company is useful. But that is not re engineering. That is not re engineering because the reason we have jobs at all is because of outcomes we're trying to achieve. The reason we have sales jobs, marketing jobs, HR jobs, finance jobs, operational jobs in stores, nurses, et cetera, is, is to do work. And knowing that some of the tasks can be automated doesn't tell you the relationship between the tasks and the work. You could assume perhaps that a lot of these tasks are really operational routine and we should automate them. But maybe not. Maybe one of the tasks, if you're a nurse, has to do with checking with the head nurse to look something up in a database to validate if a drug is safe. It's going to look like a task that should be automated by AI. But that's not a very small automation. That's a very Strategic automation. So you're not going to know what this task intelligence thing can do for you until you sit down and work with the business itself on the workflow you're trying to automate, which means you're really doing business consulting, not job task consulting. Now this reminds me a lot of the skills based organization stuff we just did because HR people love to do projects like this because they sound so useful and they are useful. I mean, I'm not saying they aren't, but you won't get that much out of it until you sit down and look at the outcome. And my opinion on this, and you'll be reading more about this in January, is that the re engineering that's going to have the highest ROI is not going to be job task analysis, it's going to be business analysis. It's going to be looking at the technology and seeing at an end to end level what it can do and then re engineering a brand new process around it that has deeper customer intimacy, faster time to market, higher quality, better data, et cetera. So we talked about that and I tried to sort of describe that for people. I really want to recommend you read our book Superpowered. It is completely 100% relevant to all of these issues you're facing right now. And it gives you you this in great detail at a business level. And you can give it to your CEO or your CFO or other people in the company if they, if they think the purpose of AI is to eliminate a bunch of jobs. I beg to differ and I will have that debate. Or we'll do a workshop with any one of you guys and I'll prove it to you. Okay, so that comes up a lot. The third thing that comes up a lot is, okay, all of these tools. Seventeen agents from Workday, a thousand agents from Oracle, hundreds of agents from SAP. Where do we start? So this came up in a lot of the meetings that I had in Europe, including All over Workday Rising. Well, there are four major ROI areas of AI today and these are not going to go away. [00:17:07] They're all systemic and they all fit into HR 2030. The first is the employee experience. I'm sure almost all of you are trying to figure out how to build a chatbot like agent for employees so they don't have to keep asking questions about their days off, their schedule, their training, their 401k, whatever that is. Poorly named Employee Self Service. [00:17:34] I think it should be renamed Employee Enablement because it's much bigger than self service because we're not talking about Just self service. We're talking about empowerment. [00:17:45] I'm an hourly worker at a Starbucks. I got 10 minutes off, I want to browse around and look for a new shift to make more money and how do I get promoted? Where's that data? Where do, what system do I go to for that? That's not employee self service. To look up the payroll and there's thousands and thousands of use cases. That is a big project. [00:18:07] It's huge. It's central to what you do in HR. [00:18:11] And you can start there, you can look at ServiceNow, you can look at what work they've done with Sana concierge, you can use Galileo to build a digital twin. There's all sorts of cool things you can do there. That's number one. Number two is in recruiting. If you're doing a high volume of recruiting and almost every company is, and you want to integrate recruiting with talent mobility and you want to be more strategic about it and you want to eliminate the number of tools and you want to start outsourcing or stop outsourcing or look at the relationship between recruiting and the use of agents internally and you want to have a better workforce planning system. [00:18:44] All of that is AI opportunity. I won't go through the vendors landscape there, but highly mature assessments, better data about high performers, faster screening, and yes, workers hate being interviewed by AI agents. So I'm not saying you have to do AI interviewing to do all this, but there's lots and lots of evolution there. And unfortunately the vendor landscape here is a mess because a lot of the vendors started in this pregen AI, so they're kind of stuck trying to catch up. But that's number two and there's out of the box solutions from Workday, from SAP, from others to do that. Number three, an area that's not getting nearly enough attention is corporate training and learning. [00:19:23] Massive, massive opportunity. $400 billion market, every company does it. It goes the range from onboarding to sales training to technology IT training to leadership development to operational training to safety training to compliance training to channel training. All of that is old paradigm boat anchor kind of stuff where you have lots of legacy technology and legacy content can be completely re engineered. We will be happy to help any of you with this. We've helped a lot of companies do this now and we've written a whole definitive guide on dynamic enablement. That's another massive area. The fourth that I like to discuss relative to AI, and there's many, many others, is the enablement of business partners. Now if you listen to my HR 2030 little narrative that I did in the webinar the other day. The business partner role is a very important role. There's no way HR is going to exist over the time without embedded professionals in the business. You can't sit around in corporate and do all your HR in the corporate tower. You can't. So you've got people working with business units on various performance recruiting, leadership training, operational or design issues. Those people need data and information and tools to do their job well. They need to be skilled. We do a lot of HR business partner upskilling and they need access to information so they don't waste all their time searching for stuff. There is a big opportunity and this is coupled to the employee experience stuff to give them an expert agent to help them. We've built such a thing with Galileo and you can take the Galileo digital twin and you can add to it your own data and you essentially have an HR business partner expert to help HR business partners. In fact, one of the projects we're working on with a couple of clients is to take the existing business partner population, take the top two or three gurus who everybody respects the most, who are the most senior, most experienced, and use the AI to capture their knowledge using documents and emails and other forms of document capture and data capture, and build a synthetic digital twin That's a digital HR business partner. ServiceNow is working on this, we're working with them, we're going to do an embed project with them. [00:21:43] And I think it's extremely valuable because that digital twin, or whatever we call it, digital HR business partner is a point of architecture in the HR 2030 architecture that delivers your strategies and solutions directly to senior leaders and executives. If you listen to IBM talk about their journey in hr, which has gone on for more than a decade, their HR business partner digital twin is a very strategic part of their solution because they made a decision a long time ago to get rid of the number of business partners for lower level managers and only provide it to high level managers. Partly for cost reasons, primarily not for strategy reasons. But the reason this is important is you build essentially a leverage point for every other thing you do in the company to reach people quickly with this infrastructure. So there's four huge areas to work on. And what we do with companies directly, either through workday or directly through us, is is we will do a workshop with you for it takes a day or two and we will walk you through this, we will show you how these things work, we will give you demos and then we will Work with you as a team within hr, with business leadership involved, to prioritize where you put your AI investments. Okay, last topic, a little bit is about the culture of your company and the initiatives you embark on relative to AI. So in the AI industry we have basically a culture of greed. I don't think I've ever seen it at this level of intensity. But everybody is only interested in making as much money as they can. And that is fueled by the stock market, by Milton Friedman, by Trump, by a whole bunch of money grabbing interests and in the economy. [00:23:37] So they're building stuff as fast as they can, trying to charge as much as they possibly can, and not thinking that much about solutions, but really focusing on the land grab that's happening in the marketplace on your side. As a corporate person, I think we have the opposite culture needed. We're not really interested in doing AI for its own sake. We, we're not interested in just laying a bunch of people off so Wall street looks good. We want to make our companies better. We want to take the business we're in, healthcare, oil and gas, transportation, retail distribution, airlines, whatever it is, and we want to just do more of it at a higher quality, at a lower cost. So we see AI and we should see AI not as a technology that's going to revolutionize everything all at once, but an amazing new tool set that will embark on replacing and updating all of our current infrastructure to do the things that we think are valuable, better and better and better. And this is really the theme of the 2027 imperatives. And so the culture that we want to build is not radical layoffs of everybody with agents running around hundreds at a time, but slowly but deliberately experimenting, building an AI lab, learning how to build agents, adopting technology that we like, possibly replacing technology that becomes obsolete, becoming savvy about the features, understanding multi agent systems, by the way, which is huge. Nobody's really worked on this, but we're working on it and then doing projects that have high alignment with our business strategy and our current market opportunities. That has nothing to do with a vendor convincing us that we're going to save 30% of our tasks in XYZ in area just because they think it's good for them. So I really think you have to absorb the vendor market, internalize it. Let us help you with this and then take a step back and decide where to focus. And this is where the high ROI projects are. We've looked at a lot of them. There are lots of obvious business needs that companies have that can be revolutionized by AI. And that's a culture of internal development, a culture of redeployment, a culture of growth, not layoffs, a culture of talent density, a culture of new models of leadership, super managers and a culture of positive expectations for a technology that is very immature. And let's face it, this stuff is very immature. Whatever Sam Altman or Dario Amadai or Satya Nadella says, even they don't know how to use it very well. I mean, I love Satya Nadella, he's one of my favorite executives. But even within Microsoft, a lot of their AI projects haven't gone well. And so let's learn how to experiment and focus these tools through the education of our own people on this new era before us. It is the super worker era. It is the super powered organization. [00:26:47] All of these things we write about are absolutely happening right before our eyes. But we have to accept the fact that this is happening in a market of greed, immaturity, power grabbing and incumbents. And if you sort of take a step back and just think about that, you're going to make very sound decisions. [00:27:07] I'm really excited about the year ahead. These last two weeks gave me a very, very deep dose of reality. I want to thank all of you that I met with while I was out on the road and all of the wonderful warm wishes and greetings and meetings that we had. There will be a lot more of this to come next year. Have a great holiday Thanksgiving for those in the US and stay tuned for more to come. Bye for now.

Other Episodes

Episode 0

July 17, 2021 • 00:21:01
Episode Cover

The SEC Human Capital Rules And Why Government Regulations Can Be Good

In this podcast I discuss the new SEC Human Capital Disclosure rules, why they’re important, and how government regulation as a whole is a...

Listen

Episode 0

February 21, 2026 • 00:21:33
Episode Cover

A Sneak Peek Under The Covers of AI-Fueled Recruiting, And Lots More

This week I explain some of the very cool things going on in AI-Fueled Recruiting (hot space), and also discuss how to start integrating...

Listen

Episode 0

June 05, 2021 • 00:13:44
Episode Cover

Under The Covers Of The Josh Bersin Academy. Now A Business Imperative.

In this podcast, I go under the covers of JBA, The World’s Professional Development Academy for HR. After two years, forged by the Pandemic,...

Listen